6/20 Run the second version of simulation code, one of the output file "event_I_Q_1Mn.data_perAS.txt" is not generated.

6/21 Output "event_I_Q_S1_1Mn.data_queryLatency.txt" is generated with not decent format. Try to find out why the other output file is lost. Run run_1M needs near 20 hours.

6/22 The last empty line in event data input files ruins everything. This is the reason why output file "event_data_perAS.txt" is missing. Now the simulation runs.

6/27 Write eventGeneration file to generate event list including insert, update and query. In update event and query event, GUID are divided into three groups with different percentage events in whole events. The next step is to adjust simulation code to get latency data about all kinds of events.

---

In this event generation list, 1000 GUID generate 1000 updates. Group 1 ( 20% GUID, from 1 to 200) generate 40% of the total update. Group 2 ( 40% GUID, from 201 to 600) generate 30% of the total update. Group 3 ( 40% GUID, from 601 to 1000) generate 30% of the total update.

---

7/14 query entry format: | src GUID | scr AS | dest GUID |

The generation of dest GUID is based on popularity property.

query group: web contents, voice call, others

1. web group: static group, mobile group 1.1 traffic between static and mobile group: popularity distribution is zipf like law 1.2 traffic within mobile group: similar to peer-to-peer network